VOGONS


First post, by marshallz

User metadata
Rank Newbie
Rank
Newbie

I am trying to find more info on this card (jumper settings,etc) but am struggling to even identify it. The part number P/N:9301-01 yields no search results, while searches including other chipsets such the promise controller, don’t seem to have any combination with a Gd5426.

It has a CL-GD5426 VGA, a Promise PDC20230C controller and a Winbond W83757F chip. I am hoping somebody may recognise the manufacturer as a start by the PN format, or SN which references “DC”. Did any other manufacturers of that time use similar yellow SN stickers?

Thanks!

pmgqBUGij

Attachments

  • Card.jpg
    Filename
    Card.jpg
    File size
    272.88 KiB
    Views
    582 views
    File license
    Public domain

Reply 2 of 10, by marshallz

User metadata
Rank Newbie
Rank
Newbie
Doornkaat wrote on 2023-02-01, 11:12:

Are there any additional hints on the backside?

Unfortunately, no. There are no markings on the PCB, no FCC ID. On the back side there is only a basic sticker with a "TP" logo and a serial number which is different to the front sides serial number. I beleive it is just for the company who assembled the machine it was originally in for warranty purposes.

Reply 3 of 10, by douglar

User metadata
Rank Oldbie
Rank
Oldbie

Looks similar to the Octek VL-COMBO, although the part number formats don't look similar, so it's probably not Octek. Re: BIOS needed for Octek VL-COMBO

Cards from this era should have an FCCID that identify the manufacturer if you look them up on the FCC web site.

Check the back to see if it is there.

Reply 4 of 10, by marshallz

User metadata
Rank Newbie
Rank
Newbie
douglar wrote on 2023-02-01, 15:11:
Looks like a slightly older version of the Octek VL-COMBO […]
Show full quote

Looks like a slightly older version of the Octek VL-COMBO

Re: BIOS needed for Octek VL-COMBO

Cards from this era should have an FCCID that identify the manufacturer if you look them up on the FCC web site.

Check the back to see if it is there.

The Octek combos appear to be the most common of the VL combo cards. But, from what i just read in the post from the link.. if mine were an Octek it would have to be > rev 2.1 and < rev 3.2.
There is definitely no FCCID. There is no copyrights, no Octek logo and no revison number anywhere. It is a very odd card!

Reply 5 of 10, by douglar

User metadata
Rank Oldbie
Rank
Oldbie

And the octek cards use a different floppy IO controller too, now that I look at it.

Sometimes the FCCID can be on a sticker that falls off. Or maybe this was never intended for the US.

Can you dump the VGA bios? There might be a reference to the OEM in there somewhere.

Reply 6 of 10, by mkarcher

User metadata
Rank l33t
Rank
l33t
marshallz wrote on 2023-02-01, 10:43:

I am trying to find more info on this card (jumper settings,etc) but am struggling to even identify it.

You don't necessarily have to find info about the jumpers, you can reverse engineer it yourself without too much hassle. There are no jumpers next to the graphics chip, so that function is likely unjumperable. There are three jumpers next to the VL IDE interface. I expect the single one to be either a 1F0/170 base address switch or an enable/disable jumper, and the paired jumper to select one of four IDE communication speeds. The available speeds are the same for all PDC20230 cards, so the documentation for another one should tell you the actual rates (likely given as number of bus cycles). You can use the "buffered disk read" number from speedsys to find out the general performance of the four possiblities, and map them to the specified speeds.

So only the big bunch of jumpers near the W83757F remain. The datasheet for the W83757F is available at alldatasheet.com for example. It explains six jumpers used to set the base addresses. You can either buzz them out using a continuity tester or just experiment what jumper setting causes which base addresses to show up in the BIOS system configuration screen. Another set of jumpers is for the IRQs: The W83757F has three IRQ output pins: One for the parallel port and one for each serial port. These output pins are routed to IRQ5/IRQ7 for the parallel port (likely a single 3-pin jumper, JP1 would be a prime candidate for that), and the COM ports are often jumperable to IRQ2/3/4/5 with a jumper that can be set in one of four positions. These jumpers just connect ISA IRQ pins to the specific W83757F pins, so buzzing them out as easy, again.

Reply 7 of 10, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

There's some CL boards identified as Gainward on VGA Legacy that appear to have the same serial number sticker and oval QC OK sticker, but those also seem a bit generic.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 8 of 10, by douglar

User metadata
Rank Oldbie
Rank
Oldbie
BitWrangler wrote on 2023-02-01, 20:08:

There's some CL boards identified as Gainward on VGA Legacy that appear to have the same serial number sticker and oval QC OK sticker, but those also seem a bit generic.

You have a 9301-01 card. Looks like SquallStrife had a 9301-10 card one once upon a time:
Re: "New" Goodies! (Cirrus Logic "Two-in-one" within!)

They determined that it was a Gainward, but Gainward wouldn't admit to it.

Seems like the SN number sticker may have been applied over the FCCID on your card.

Reply 9 of 10, by marshallz

User metadata
Rank Newbie
Rank
Newbie
douglar wrote on 2023-02-01, 18:02:

And the octek cards use a different floppy IO controller too, now that I look at it.

Sometimes the FCCID can be on a sticker that falls off. Or maybe this was never intended for the US.

Can you dump the VGA bios? There might be a reference to the OEM in there somewhere.

Unfortunately, there are no stickers left to fall off. I am 100% it is made by Gainward/Cardex.

mkarcher wrote on 2023-02-01, 18:37:
marshallz wrote on 2023-02-01, 10:43:

I am trying to find more info on this card (jumper settings,etc) but am struggling to even identify it.

You don't necessarily have to find info about the jumpers, you can reverse engineer it yourself without too much hassle. There are no jumpers next to the graphics chip, so that function is likely unjumperable. There are three jumpers next to the VL IDE interface. I expect the single one to be either a 1F0/170 base address switch or an enable/disable jumper, and the paired jumper to select one of four IDE communication speeds. The available speeds are the same for all PDC20230 cards, so the documentation for another one should tell you the actual rates (likely given as number of bus cycles). You can use the "buffered disk read" number from speedsys to find out the general performance of the four possiblities, and map them to the specified speeds.

So only the big bunch of jumpers near the W83757F remain. The datasheet for the W83757F is available at alldatasheet.com for example. It explains six jumpers used to set the base addresses. You can either buzz them out using a continuity tester or just experiment what jumper setting causes which base addresses to show up in the BIOS system configuration screen. Another set of jumpers is for the IRQs: The W83757F has three IRQ output pins: One for the parallel port and one for each serial port. These output pins are routed to IRQ5/IRQ7 for the parallel port (likely a single 3-pin jumper, JP1 would be a prime candidate for that), and the COM ports are often jumperable to IRQ2/3/4/5 with a jumper that can be set in one of four positions. These jumpers just connect ISA IRQ pins to the specific W83757F pins, so buzzing them out as easy, again.

It is looking more and more likely that this may be the route i will need to go down. I was trying to find a controller card with the same Winbond chipset in hopes of being able to cheat at this, thinking it may have the same jumper block but they oddly seemed to be a 1-2-3 pin setting, whereas my card is a 1-2 pin setting Open/Close. The only CARDEX I/O controller i could find had a different jumper block, too.

douglar wrote on 2023-02-02, 03:06:
You have a 9301-01 card. Looks like SquallStrife had a 9301-10 card one once upon a time: Re: "New" Goodies! (Cirrus Logic "T […]
Show full quote
BitWrangler wrote on 2023-02-01, 20:08:

There's some CL boards identified as Gainward on VGA Legacy that appear to have the same serial number sticker and oval QC OK sticker, but those also seem a bit generic.

You have a 9301-01 card. Looks like SquallStrife had a 9301-10 card one once upon a time:
Re: "New" Goodies! (Cirrus Logic "Two-in-one" within!)

They determined that it was a Gainward, but Gainward wouldn't admit to it.

Seems like the SN number sticker may have been applied over the FCCID on your card.

Yes, I was pointed to that same post from a Facebook group member and i could not believe the similarities! the -10 appears to have less circuitry but same jumper blocks. The PCB layout is slightly different toward the COM IDC connectors. I removed the serial number sticker on mine and oddly enough, there is no FCCID under it either!

I am 100% sure that both cards are made by the same manufacturer. The FCCID of the 9301-10 card says the card is made by Gainward. Gainward was founded in 1984. Their older websites reference a "CARDEX" line. Cardex was the North American division of Gainward formed in 1991. The Cardex website seemed to be absorbed after 1996 into the Gainward site. I think i read Gainward is now owned by Palit? Could be wrong there... All Cardex/Gainward sites reference nothing less than a Cirrus Logic 5440 which I think was released in 1995. It is as if they wanted to forget any older devices prior to that date! None of their sites seem to reference anything other than graphics cards, eventhough Cardex I/O controller cards by themselves exist!

Never in my life have I had so much difficulty identifying an expansion card! Just for laughs, i may try and email them. I think the author of the other post originally posted that he emailed Gainward back in 2011. I can only imagine the response i would get today!

Reply 10 of 10, by mkarcher

User metadata
Rank l33t
Rank
l33t
marshallz wrote on 2023-02-02, 14:43:

It is looking more and more likely that this may be the route i will need to go down. I was trying to find a controller card with the same Winbond chipset in hopes of being able to cheat at this, thinking it may have the same jumper block but they oddly seemed to be a 1-2-3 pin setting, whereas my card is a 1-2 pin setting Open/Close.

1-2-3 configuration jumpers usually have +5V on 1, GND on 3 (or the other way around), and the configurable pin on 2. This allows to pull the pin to either GND or +5V using 1-2 or 2-3. If the jumper is just open/close, most of the times "close" pulls the configurable pin to GND, and there is a resistor to +5V (a "pull up" resistor) that pulls the configurable pin to +5V in case the jumper is open. If everything else is equal, you should find a 1:1 mapping between open/close on your card and 1-2/2-3 on the other cards. For the W83757F, this applies to just 6 of the jumpers, as only the base address setting is configured using pull-up/pull-down strap pins. IRQs are routed completely externally. I guess in your case matching to different cards is not easier than just probing your card.